At GitHub, we’re committed to keeping the GraphQL API dependable for everyone. To reinforce the protection of our infrastructure, we’re introducing new safeguards to limit resource consumption.

What kind of limits?

While we can’t share all specifics to preserve the integrity of our platform, note that these limits are different from rate limiters. They don’t track how many requests you send or how quickly. Instead, they cap the execution resources a single query can consume. Patterns like the following may cause queries to hit these resource limits and return partial results with errors:

  • Requesting very large numbers of objects or deeply nested relationships in a single query
  • Using large first or last arguments across multiple connections at the same time
  • Fetching exhaustive details for each object, such as all comments, reactions, and related issues for every repository

How this affects your integration

In typical usage, you shouldn’t notice any change. However, if your integration issues expensive queries, you may see more partial responses accompanied by errors indicating that resource limits were exceeded.

Learn more about the GraphQL API’s rate and query limits at GitHub.